How Big a Problem Is Uber’s Self-Driving-Car Crash for the Autonomous-Vehicle Trend?

In this segment from the MarketFoolery podcast, host Chris Hill is joined by Motley Fool Asset Management's Bill Barker to talk about the latest less-than ideal news in autonomous vehicles -- the fatal accident involving an Uber test autonomous vehicle and a pedestrian. All of its tests are on hold while the accident is investigated. It's the first such death, and it will have an effect on every company in the space. But the Fools don't necessarily see it as changing the path toward self-driving cars.

A full transcript follows the video.

This video was recorded on March 20, 2018.

Chris Hill: Let's move on to Uber, which has suspended its self-driving-car initiative that it had been testing after an autonomous Uber car -- which by the way, had an emergency backup driver behind the wheel -- this vehicle struck and killed a woman on the street in Tempe, Arizona. Uber had been testing this initiative in Tempe, in Pittsburgh, in San Francisco, in Toronto. All of those have been suspended.

I think, obviously, the spotlight is on Uber, and rightly so. But I think this is one of those stories that affects absolutely every company that is looking to get into the autonomous driving space.

Bill Barker: Yeah, and it'll be a big story for today. There's a fatality. But over time, people will grow accustomed to what the potential rewards are for getting the technology right. There are about 100 fatalities a day on the road in the U.S. This is the first autonomous-vehicle death, and I'm sure it won't be the last. I'm sure, ultimately, there will be thousands, because you're talking about a country where we are used to and accept some 30,000-35,000 deaths a year through the mistakes of human drivers.

And that's actually a much better number than it used to be. We have gotten better, car makers have gotten better, drivers have gotten better about not drinking and driving, and things are safer. And yet, still, 35,000 deaths a year. And I think we all know the dangers when we drive. This seems to be scary because there's no driver. As you said, there was a driver behind the wheel, on top of the autonomous system. So, could this have been an accident that would have occurred if the driver had been fully engaged? I don't know. I guess they're going to investigate that and report back. I'm sure the autonomous vehicle community is hoping that this was an accident that's not completely the responsibility of the system, but just an accident which happens because of pedestrian mistake. I don't know. That's what will come out eventually.

Hill: I'm glad you mentioned the annual stats because that's important context to have. And yet -- I'll just speak for myself. I think this is an unfair expectation on my part, but I don't think I'm alone in this, that fairly or unfairly, and it's probably unfairly, my expectation is that, if we're going to have autonomous vehicles, my expectation is that they'd be perfect. And I realize that's not a reasonable one, but it's far more an emotional one than an intellectual one.

So, we can look at a future 10 years from now, 20 years from now where the number of deaths on an annual basis from autonomous vehicles is in the low hundreds, as opposed to in the tens of thousands that we currently have. And yet, it still sort of feels like, "Oh, I don't know ... " Maybe it's because we're in the early part of this process, and not 20-50 years in the future, when of course we'll to look back -- in the same way now that we look back on driving in the 1950s and the 1960s and we think, it's insane that there weren't seatbelts and we didn't have seats for kids, and drinking and driving was just like, "Yeah, sure, of course we're drinking and driving."

Barker: I think, it would be nice if this was the kind of technology that could come out and be perfect from the beginning. But it's not going to be. Air travel, it would be great if there were zero fatalities from that, but there never have been. The stats get better and better over time, and that's what one should expect. If you can step back from the tragedy of a single death, there will be additional tragedies. Hopefully there will be significantly fewer than the number of tragedies we are used to accepting from the current system, which is already much better than it used to be. Things get better and better. This will be part, one would hope, of improvement, but it'll never be perfect.

Hill: Are you surprised at that list of cities where this is being tested? It's Tempe, Pittsburgh, San Francisco and Toronto. Really, what I'm surprised by is Pittsburgh, San Francisco and Toronto. I just assumed that, "We're going to test this not in a closed course, we're going to test autonomous vehicles in cities. Let's start with smaller cities where the weather isn't really an issue." So, the fact that it's being tested in places like Pittsburgh and Toronto, in terms of winter weather, San Francisco, which is an insane place to drive no matter what the weather is ... I don't know, I'm just surprised by that.

Barker: Hopefully they're not testing the treacherous weather conditions as part of this at the moment. Maybe they are. But, you'd certainly want to be doing that on a closed course, because that's just one more variable. You don't need to test multiple variables at the same time this early in the process. You can be testing in Pittsburgh and just not test it on days when the weather, even if it's just raining, you can pull that day as a day of testing. You're not testing with no driver behind the wheel, I think, I don't know. I haven't been up to date on all the different--

Hill: That's ultimately the goal, though.

Barker: That's ultimately the goal, but you don't test that part until you have a lot of data, and I don't know that municipalities are going to be approving such tests for a while.